Quantification of splice variants using real-time PCR

Abstract

A reliable and robust method for measuring the expression of alternatively spliced transcripts is an important step in investigating the significance of each variant. So far, accurate quantification of splice variants has been laborious and difficult due to the intrinsic limitations of conventional methods. The many advantages of real-time PCR have made this technique attractive to study its application in quantification of splice isoforms. We use skipping of exon 37 in the NF1 gene as a model to compare and evaluate the different strategies for quantitating splice variants using real-time PCR. An overview of three different possibilities for detecting alternative transcripts is given. We propose the use of a boundary-spanning primer to quantify isoforms that differ greatly in abundance. We describe here a novel method for creating a reliable standard curve using one plasmid containing both alternative transcripts. In addition, we validate the use of an absolute standard curve based on a dilution series of fluorometrically quantified PCR products.

Figure 1

Schematic representation of the pUC18NF1+37-NF1Δ37 plasmid, which contains both alternative transcripts. ORI, origin of replication; AP^r, ampicilin resistance gene. Two standard curves are generated from the same serial dilutions, thus providing complete equality of both curves.

Figure 2

Overview of different possibilities to detect and distinguish alternative splice variants. (A) Detection by a boundary spanning probe. (B) Detection by a boundary spanning primer. (C) Quantification by subtraction, only suitable for equally abundant transcripts.

Figure 3

Real-time PCR quantification for each dilution point of the pUC18NF1+37-NF1Δ37 plasmid amplified by primer pair 1 (grey) and 2 (black). Equal quantities are observed for each tested dilution point of both standard curves.

Figure 4

Quantification of NF1Δ37 using a boundary spanning probe (A) or a boundary spanning primer (B). (A) Equal amounts of water (red curves) or NF1+37 (blue curves) were added. Equal C_tvalues are obtained for the first three dilutions. Influence on amplification efficiency from the presence of full-length transcript is already observed for dilution 4 by a reduction of end-point fluorescence and an increase in C_tvalue for the blue curve. At higher C_tvalues quantification is no longer reliable as can be seen for the C_t difference between dilutions 5. (B) No difference is seen between the series with the addition of water (red) and NF1+37 (blue).

Figure 5

Representative amplification plots of EBV (blue), fetal liver (red) and cerebellum (green). a-curves represent NF1+37, b-curves NF1Δ37.

Figure 6

Logarithmic scatterplot of quantities of serial dilutions of PCR-NF1+37 (x-axis) and PCR-NF1Δ37′ (y-axis) products measured by a standard curve based on a dilution series of pUC18NF1+37-NF1Δ37 plasmid DNA.